mobilenetv3-HandwritingStrip-3class-v1

This model is a fine-tuned version of on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 0.0135
• Accuracy: 0.9952
• Precision: 0.9960
• Recall: 0.9922
• F1: 0.9940
• Roc Auc: 0.9999

Model description

More information needed

Intended uses & limitations

More information needed

Training and evaluation data

More information needed

Training procedure

Training hyperparameters

The following hyperparameters were used during training:

• learning_rate: 0.0003
• train_batch_size: 64
• eval_batch_size: 64
• seed: 42
• optimizer: Use OptimizerNames.ADAMW_TORCH_FUSED with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: linear
• num_epochs: 3
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	Precision	Recall	F1	Roc Auc
0.4220	0.0847	10	0.5832	0.7518	0.7690	0.7867	0.7564	0.9486
0.0928	0.1695	20	0.1893	0.9379	0.9250	0.9440	0.9324	0.9908
0.1184	0.2542	30	0.1775	0.9356	0.9194	0.9436	0.9262	0.9944
0.1044	0.3390	40	0.0801	0.9690	0.9587	0.9680	0.9629	0.9980
0.0729	0.4237	50	0.0806	0.9714	0.9607	0.9701	0.9649	0.9973
0.0726	0.5085	60	0.0985	0.9547	0.9395	0.9579	0.9464	0.9979
0.0501	0.5932	70	0.0704	0.9761	0.9656	0.9779	0.9709	0.9982
0.0419	0.6780	80	0.0422	0.9881	0.9828	0.9880	0.9853	0.9993
0.0452	0.7627	90	0.0256	0.9928	0.9920	0.9901	0.9911	0.9998
0.0733	0.8475	100	0.0515	0.9785	0.9689	0.9799	0.9738	0.9991
0.0617	0.9322	110	0.0492	0.9809	0.9846	0.9686	0.9756	0.9997
0.0250	1.0169	120	0.0247	0.9881	0.9843	0.9861	0.9852	0.9997
0.0367	1.1017	130	0.0248	0.9881	0.9843	0.9861	0.9852	0.9998
0.0477	1.1864	140	0.0237	0.9881	0.9861	0.9861	0.9861	0.9997
0.0446	1.2712	150	0.1330	0.9547	0.9389	0.9597	0.9459	0.9983
0.0428	1.3559	160	0.0419	0.9809	0.9747	0.9781	0.9764	0.9994
0.0267	1.4407	170	0.0281	0.9857	0.9792	0.9860	0.9824	0.9997
0.0243	1.5254	180	0.0172	0.9952	0.9941	0.9941	0.9941	1.0000
0.0220	1.6102	190	0.0172	0.9928	0.9921	0.9921	0.9921	0.9999
0.0338	1.6949	200	0.0213	0.9905	0.9882	0.9901	0.9891	0.9999
0.0171	1.7797	210	0.0177	0.9928	0.9902	0.9920	0.9911	0.9999
0.0304	1.8644	220	0.0168	0.9928	0.9920	0.9901	0.9911	0.9999
0.0200	1.9492	230	0.0191	0.9905	0.9900	0.9862	0.9880	0.9999
0.0279	2.0339	240	0.0177	0.9905	0.9900	0.9862	0.9880	0.9999
0.0213	2.1186	250	0.0171	0.9905	0.9900	0.9862	0.9880	0.9999
0.0058	2.2034	260	0.0238	0.9881	0.9843	0.9861	0.9852	0.9998
0.0104	2.2881	270	0.0275	0.9881	0.9843	0.9861	0.9852	0.9998
0.0130	2.3729	280	0.0173	0.9881	0.9843	0.9861	0.9852	0.9998
0.0047	2.4576	290	0.0153	0.9928	0.9940	0.9882	0.9910	0.9999
0.0096	2.5424	300	0.0160	0.9928	0.9940	0.9882	0.9910	0.9999
0.0139	2.6271	310	0.0155	0.9928	0.9940	0.9882	0.9910	0.9999
0.0129	2.7119	320	0.0149	0.9928	0.9940	0.9882	0.9910	0.9999
0.0106	2.7966	330	0.0143	0.9928	0.9940	0.9882	0.9910	0.9999
0.0136	2.8814	340	0.0137	0.9952	0.9960	0.9922	0.9940	0.9999
0.0194	2.9661	350	0.0135	0.9952	0.9960	0.9922	0.9940	0.9999

Framework versions

• Transformers 5.3.0
• Pytorch 2.10.0+cu128
• Datasets 4.0.0
• Tokenizers 0.22.2